Investigation into the physical stability of a eugenol nanoemulsion in the presence of a high content of triglyceride

Abstract

The physical stability of a eugenol (Eu) nanoemulsion in the presence of long-chain and medium-chain triglycerides with a high incorporation ratio (30 wt% to 80 wt% in the lipid phase) was investigated. The stability of the nanoemulsion was evaluated by dynamic light scattering and a centrifugal-force-based LUMi-Sizer. With an increase in the content of triglycerides, the particle size of the nanoemulsions decreased first and then increased. The variation pattern of the particle size in the presence of different contents of triglycerides formed a V-type shape. The presence of 60 wt% triglycerides resulted in the smallest particle size, which continued increasing within 28 days of storage. LUMi-Sizer analyses demonstrated that the instability of the nanoemulsion caused by the addition of 30–50 wt% of triglyceride was due to sedimentation of emulsion droplets. On the other hand, the instability caused by 60–80 wt% of triglyceride was due to the floating of emulsion droplets. The interfacial tension variation had the same trend as the particle size variation at the corresponding content of triglycerides. Thus the interfacial tension might be responsible for the triglyceride-caused instability of the Eu nanoemulsion.

---

1. Introduction

Eugenol (Eu) is a well-studied essential oil and belongs to a class of volatile phytophenols derived from clove, cinnamon and bay leaves. This compound exhibits excellent antimicrobial activity^1–3 and is generally recognized as safe.⁴ However, its application is limited due to its high volatility and poor water-solubility. An oil-in-water emulsion is a carrier system and commonly used to encapsulate functional lipophilic compounds such as nutraceuticals, antioxidants, flavors, pigments and antimicrobials in food, cosmetic and pharmaceutical industries.^5,6 It can improve their solubility and retard their volatility.^7,8 Unfortunately, it has been reported in the literature that the incorporation of essential oils (such as Eu, orange oil and thyme oil) weakens the stability of emulsions against ripening or coalescence due to the relative high water-solubility, low molecular weight and high movability of essential oils.^9–13 Ostwald ripening is found to be the main mechanism that destabilizes essential oil-loaded nanoemulsions. It is attributed to diffusion of the dispersed phase through the continuous phase and a spontaneous trend toward a minimal interfacial area between the continuous phase and the dispersed one.^14,15 The occurrence of Ostwald ripening causes the growth of larger particles at the expense of smaller ones, resulting in the instability of emulsions.

Decreasing the water-solubility of the organic phase is a common strategy to suppress Ostwald ripening. Highly hydrophobic substances (ripening inhibitors), such as long-chain triglycerides (LCTs), medium-chain triglyceride (MCTs) or alkanes, are usually incorporated into the organic phase.⁹ It has been reported that incorporation of ≥60 wt% corn oil (LCT) and 50 wt% MCT inhibited the Ostwald ripening of a thyme oil emulsion.⁹ A stable peppermint oil nanoemulsion could be obtained by adding MCT into the oil phase.¹⁶ These results showed that the particle size of the essential oil-loaded emulsion decreased and then slightly increased to a plateau with the increasing concentration of triglycerides. Our previous research also demonstrated that incorporation of bean oil (BO, >20 wt% in the lipid phase) can improve the stability of Eu nanoemulsions when sodium dodecyl sulfate (SDS) was used as the surfactant.¹³ However, when Tween80 (T80) was used as the surfactant, a high amount of BO caused a decrease in the second particle size and the worse stability of Eu nanoemulsions, which has seldom been reported previously. Hereby, we hypothesized that the selection of surfactant should be taken into account when ripening inhibitors were used to suppress the occurrence of Ostwald ripening.

The presented study aimed to evaluate and confirm the physical stability of an Eu nanoemulsion using T80 as surfactant in the presence of a high incorporation ratio of triglyceride. Two techniques, including dynamic light scattering based on Brownian movement and a LUMi-Sizer based on centrifugal forces, were utilized to further evaluate the physical stability of Eu nanoemulsions. The underlying mechanism of effects caused by high amounts of triglyceride on the physical stability was also investigated.

2. Materials and methods

2.1 Materials

Eugenol (Eu, >99% GR assay) was supplied by Aladdin Reagent Database Co. (Shanghai, China). BO was purchased from a local supermarket without purification. MCT was purchased from Boxing Chemical Reagent Co. Ltd (Wuhan, China). T80 was obtained from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China). Ultra-pure water with a resistance of 18.2 MΩ cm was used to prepare aqueous solutions.

2.2 Emulsion preparation

Emulsions contained a 5 wt% lipid phase and a 95 wt% aqueous phase. The lipid phase was prepared by mixing Eu and the triglyceride. BO and MCT were applied as LCT and MCT, respectively. The fraction of LCT or MCT in the lipid phase was set at 30 wt%, 40 wt%, 50 wt%, 60 wt%, 70 wt% and 80 wt%. The aqueous phase was 1 wt% T80 solution. Coarse pre-emulsions were obtained by blending the lipid and aqueous phases with a Polytron homogenizer at 26 000 rpm for 3 min in an ice bath. The resulted pre-emulsions were then homogenized by passing through a high pressured micro-fluidizer (Micro-fluidics M-110L, Microfluidics Corp., Newton, MA, USA) at 11 000 psi for five passes. The prepared nanoemulsions were kept at 4 °C prior to analyses.

2.3 Determination of particle size

The particle size and particle size distribution of nanoemulsion droplets were characterized using dynamic light scattering (Zeta sizer Nano ZS, model ZEN 3600, Malvern Instrument, Malvern, U.K.) at a scattering angle of 173°. Each measurement was an average of 15 runs. A laser beam (658 nm) was used as the light source. To prevent multiple scattering effects, the nanoemulsion was diluted in de-ionized water to reach to a droplet concentration of 0.005% (w/v) prior to the measurement. The particle size data were reported as the intensity-weighed (“Z-average”) mean particle diameter.

The particle size of samples was also measured using a laser diffraction device (Mastersizer 2000, Malvern Instruments, MA). This instrument measured the angular dependence of the intensity of laser light (λ = 632.8 nm) scattered by a dilute particle suspension. Then the particle size distribution aligned most between theoretical predictions (Mie theory) and experimental measurements. To avoid multiple scattering effects, samples were diluted to a concentration of 0.005% (w/v) using 5 mM phosphate buffer with the same pH as samples. The refractive index of BO, MCT and Eu were 1.476, 1.448, and 1.500, respectively. The particle size measurements were reported as the surface-weighted mean diameter: d₃₂ = ∑^​n_id³_i/∑^​n_id²_i, where n_i was the number of droplets of diameter d_i.

2.4 Dynamic physical stability of the emulsion

The stability of the emulsion was evaluated with a LUMi-Sizer (L.U.M. GmbH, Germany), through which a centrifugal force was applied to facilitate the occurrence of instability such as sedimentation, flocculation, phase separation and creaming. Simultaneously, near-infrared light was used to illuminate the sample to determine the transmitted light intensity as a function of time and position over the sample cell. The results were displayed as a function of the space- and time-dependent transmission profiles over the sample cell. The operational parameters of the LUMi-Sizer were as follows: volume, 1.8 mL of dispersion; wavelength: 865 nm; centrifuge rate: 4000 rpm; time_Exp: 5000 s; time interval, 5 s; temperature, 25 °C.

2.5 Measurement of viscosity and interfacial tension

The viscosity of the oil phase was measured using a rotary viscometer (LVDV-II+P, Brookfield, USA). Interfacial tension between the oil phase and aqueous phase was determined using an optical contact angle meter (Teclis Tracker, France) equipped with an oscillating drop accessory. The experiments were carried out at 25 °C. In brief, a 1 wt% T80 solution was added to the optical glass cuvette. A syringe was filled with an oil solution and the needle was fixed. After the air was evacuated, the needle containing the oil solution was submerged into the optical glass cuvette. The location of the needle was adjusted between the light source and the high-speed charge couple device (CCD) camera. The profile data was recorded with a CCD camera and analyzed using the Laplace equation.

2.6 Statistical analysis

All experiments were performed in triplicates on freshly prepared samples. The results were then reported as averages and standard deviations of these measurements. Analysis of variance (ANOVA) analysis was further performed on the presented data.

3. Results

3.1 Effects of triglyceride content on the particle size of the Eu emulsion

It has been shown that an oil phase with lower water solubility prevents Ostwald ripening and enhances the stability of emulsion.^9,16 Our research has further illustrated that the stability of an Eu emulsion is related to the type of surfactant and the lipid compositions.¹³ Especially when BO content was higher than 30 wt%, SDS-stabilized Eu nanoemulsions remained stable and the particle size slightly increased with the BO content. However, the particle size of T80-stabilized Eu nanoemulsions decreased and then increased at the same range of BO content. With the addition of 60 wt% BO, the Eu nanoemulsions had the smallest particle size.

The physical stability of T80-stabilized Eu nanoemulsions was thus investigated at a triglyceride content in the lipid phase greater than 30 wt%. BO and MCT were selected due to their non-polarity, relatively high molecular weight, and inhibitive ability to Ostwald ripening.^10,15,17,18 Fig. 1 presented the impact of triglyceride contents on the particle size of Eu emulsions after 24 h storage. Addition of BO and MCT caused a similar trend in particle size. Overall, the particle size of the Eu nanoemulsions was below 200 nm. When the triglyceride content in the organic phase was ≥30 wt%, the triglyceride caused a V-type change in particle size. The smallest particle size appeared at 60 wt% of triglyceride, corresponding to around 110 nm and 90 nm for BO and MCT, respectively. The physical stability of these samples were further investigated by static and dynamic stability test.

Fig. 1 The impact of triglyceride contents in the organic phase on the particle size of Eu nanoemulsion droplets after 24 h.

3.2 Static physical stability during the storage of the Eu emulsion in the presence of BO

The prepared emulsified solutions were milky-white and homogenized without any creaming or separation. The particle size distribution was mono-modal after preparation (Fig. 2a). The peak of the particle size distribution shifted into smaller sizes and then larger sizes, which was in agreement withchanges in particle size displayed in Fig. 1. The impact of BO on the particle size was then monitored for 28 days (Fig. 2a). The particle size of the Eu nanoemulsions kept increasing from 100 nm to 700–800 nm during storage at 60% of BO content. When the BO content was 40 wt%, 50 wt% and 70 wt%, the particle size of the Eu nanoemulsions increased from 189 nm to 208 nm, 164 nm to 309 nm, and 114 nm to 359 nm, respectively. The addition of 30 wt% or 80 wt% BO barely impacted the size changes. The increase rate of particle size caused by the addition of BO descended as follows: 60 wt% > 70 wt% > 50 wt% > 40 wt% > 30 wt% ≈ 80 wt%.

Fig. 2 (a) Left: Particle size distribution of Eu nanoemulsions with different BO contents after preparation. Right: Particle size change of Eu nanoemulsions with different BO contents during storage determined with dynamic light scattering. (b) Particle size distribution of Eu nanoemulsions with different BO contents during storage.

The particle size of emulsions was further measured by laser diffraction. As shown in Fig. 1S(A),† the particle size of the Eu nanoemulsions increased by more than 100 nm during storage only when the BO content ranged from 50 wt% to 70 wt%. This result was similar to that measured by dynamic light scattering (Fig. 2a) and the particle sizes had a unimodal distribution (data not shown). Fig. 2b showed the unimodal particle size distribution of nanoemulsions during storage. No shift of the peak curves was observed within 28 days of storage at 30 wt% and 80 wt% of the BO content. At a BO content of 40 wt%, the peak shift was observed on the 21st day and the newly formed peak was wider. For Eu nanoemulsions with 50 wt%, 60 wt% or 70 wt% BO, the peak of the particle size distribution shifted to larger sizes with an extension of storage time as well as a wider peak.

3.3 Static physical stability during storage of the Eu emulsion in the presence of MCT

The particle size distributions of Eu nanoemulsions in the presence of MCT were also mono-modal distribution after preparation (Fig. 3a). However, the particle size variation of MCT/Eu emulsions was different from that of BO/Eu emulsions (Fig. 3a). Overall, the increase in particle size was suppressed as compared to that of the BO/Eu emulsions. A significant increase in the particle size occurred after 144 h and the addition of 70 and 80 wt% MCT had insignificant effects on the particle size with 28 days storage. When the BO content was 50 wt%, the particle size of the Eu nanoemulsions increased from 99 nm to 148 nm, which was the maximum variation of particle size. The particle size increment of the other system kept around 15–30 nm. The increase rate of particle size caused by the presence of MCT descended as follows: 50 wt% > 60 wt% > 40 wt% > 30 wt% > 70 wt% ≈ 80 wt%.

Fig. 3 (a) Left: Particle size distribution of Eu nanoemulsions with different MCT contents after preparation. Right: Particle size change of Eu nanoemulsions with different MCT contents during storage determined with dynamic light scattering. (b) Particle size distribution of Eu nanoemulsions with different MCT contents during storage.

These results were further confirmed by the measurement of laser diffraction. As shown in Fig. 1S(B),† the change of particle size varied within 10 nm. It is worth noting that in the presence of 30 wt% MCT, it was difficult to accurately measure the particle size by laser diffraction (data not shown). Fig. 3b showed the particle size distributions of Eu nanoemulsions as a function of MCT content and storage time, which remained mono-modal. The shift of peaks during storage was less than that in the presence of different BO contents. On the 28th day, the particle size distribution became broader when the MCT content were 40 wt% and 50 wt%.

3.4 Dynamic physical stability evaluation of the Eu emulsion in the presence of triglycerides

The accelerated stability based on the centrifugal force was tested by using the LUMi-Sizer to determine the long-time storage stability and the underlying mechanism of the instability.¹⁹ It can reflect the movement of emulsion droplets. Fig. 4a and b displayed the emulsion stability as a space- and time-related transmission profile over the sample length in the presence of BO and MCT, respectively. Prior to the measurement, the emulsion was homogenized and very little light was transmitted through the sample. Upon the centrifugation, the heavier phase will move to the bottom of the cells causing light transmission of the more transparent aqueous phase. While the lighter and less transparent oil phase will rise to the top, forming a cream layer. In Fig. 4a and b, the red line was the first recorded profile, which then shifted into the green line during measurements. The greater the shift of the transmission with the centrifuge time was, the lower the stability of the emulsion was. Fig. 4a showed that the maximum change of profiles happened at 60 wt% BO. The instability indices were 0.092, 0.017, 0.031, 0.699, 0.328 and 0.055, when BO contents were 30 wt%, 40 wt%, 50 wt%, 60 wt%, 70 wt% and 80 wt%, respectively. In the presence of 30–50 wt% MCT, the movement of the profile was close to that in the presence of 30–50 wt% BO (Fig. 4b). While the change of the profile in the presence of 60–80 wt% MCT was different from that of 60–80 wt% BO. The corresponding instability indices were 0.194, 0.085, 0.032, 0.232, 0.131 and 0.028, when MCT contents were 30 wt%, 40 wt%, 50 wt%, 60 wt%, 70 wt% and 80 wt%, respectively. A higher instability index usually meant a lower stability of emulsions, indicating the higher clarification velocity. The results demonstrated that 60 wt% triglyceride caused the highest instability index. Additionally, the addition of 60 wt% BO caused phase separation after centrifugation (data not shown, similar to the images in Fig. 2S†).

Fig. 4 (a) Transmission profiles of Eu nanoemulsions with different BO contents measured by the LUMi-Sizer. (b) Transmission profiles of Eu nanoemulsions with different MCT contents measured by the LUMi-Sizer.

Moreover, the shift of the transmission profile can predict the instability mechanism of emulsions. When the shift of lines appears at the bottom of tube, it indicates the movement of droplets down to the bottom, which belong to the sedimentation. When the shift of lines appears on the top of the tube, it indicates the movement of droplets up to the surface, which belong to the suspension or creaming. Fig. 4 demonstrated that incorporation of 30–50 wt% of the triglyceride in the organic phase caused sedimentation of emulsion droplets. While emulsion droplets with 60–80 wt% triglyceride tended to float up, which resulted in creaming.

3.5 Viscosity and interfacial tension of the organic phase

It is well-known that the characteristics of the organic phase will affect the properties of the emulsion. Eu is a small nonpolar molecule that has a high permeability and is therefore highly prone to Ostwald ripening.²⁰ MCT and BO are medium- and long-chain triglycerides with large nonpolar triacylglycerol molecules and they have low water solubilities. Hence, they could inhibit the increase of particle sizes of Eu nanoemulsions.

Fig. 5 showed that the viscosity of the organic phase increased with increasing triglyceride content due to the larger molecular weight, higher molecular volume and viscosity of the selected triglycerides compared to Eu. Thus it was unlikely that viscosity caused the V-type evolution of the nanoemulsion particle size. When the triglyceride content was less than 10 wt%, it would not significantly change the interfacial tension between T80 and Eu (Fig. 5). With the increased addition of triglyceride from 20 wt% to 60 wt%, the interfacial tension of Eu nanoemulsions decreased. However, the interfacial tension was enhanced with further increased addition of triglyceride from 60 wt% to 100 wt%. The curve was broken in the presence of 30 wt% or 40 wt% of BO as well as in the presence of 50 wt% of MCT due to the burst of oil droplets in the measurement of the interfacial tension. The results suggested that the change in the interfacial tension might be responsible for the V-type shape of the particle size variation in the presence of different contents of BO or MCT.

Fig. 5 (a) Viscosity and (b) interfacial tension of the oil phase solution.

4. Discussion

Addition of triglyceride into the organic phase is a common strategy to manipulate the occurrence of Ostwald ripening. Triglycerides with lower water-solubility can limit the movement of soluble oil among droplets. Most of the reported results demonstrated that addition of a ripening inhibitor would first decrease the particle size of emulsions and then maintain the particle size at a plateau¹ Fig. 2. Up to a certain amount of triglyceride, emulsions were stable. In the present study, the particle size of the Eu nanoemulsion significantly increased over storage time with the addition of 50–60 wt% BO. Phase separation even occurred in Eu nanoemulsions with 60 wt% BO after 5 month storage, displayed in Fig. 2S.† This was consistent with the phenomenon after centrifugation by the LUMi-Sizer. The instability of emulsions with high triglyceride content has seldom be reported. Compared to the effect of BO, the stability of Eu nanoemulsions with MCT addition was better. As BO is a vegetable oil extracted from the seeds of soybeans, which contains complex mixtures of long-chain triglycerides (mainly C18 fatty acids). MCT is composed of C8 and C10 fatty acids, which belong to the medium-chain triglyceride class. Due to the shorter carbon chain, MCT has a smaller molecular weight/volume, lower viscosity, hydrophobicity and polarity. Then emulsion droplets containing MCT had smaller particle sizes and lower ripening rates according to the results by Wooster.¹⁵ These findings indicate the component-dependence of emulsion stability.

When the emulsified lipid phase consists of a single component (i.e., 100 wt% Eu), the increase in particle size (r) with time (t) due to Ostwald ripening in the steady state regime is given by Taylor.²¹

here, ω is the Ostwald ripening rate, α = 2γV_m/RT (γ is the interfacial tension, V_m is the molar volume of the lipid, R is the gas constant and T is the absolute temperature, c is the solubility of the lipid in the continuous phase, and D is the translational diffusion coefficient of the solute through the continuous phase). When the emulsified lipid phase is a mixed component (i.e., 30 wt% Eu + 70 wt% BO or MCT), an oil with a relatively high solubility (soluble oil) has a greater mobility between droplets. Due to the transport of soluble oil, the particle size of droplets rich in soluble oil becomes larger over time, while that of droplets rich in insoluble oil becomes smaller. This creates an imbalance of the chemical entropy of oils. To predict the stability of the mixed oil nanoemulsion, Kabalnov's theoretical equation can be utilized.²²
where α is the same as above and X₂ is the mole fraction of insoluble oil in the emulsion. According to the equation, the stability of the emulsion can be in three regions depending on the insoluble oil molar fraction. Wooster found that with smaller amounts of insoluble oil the obtained nanoemulsion was unstable.¹⁵ A high amount of insoluble oil caused a thermodynamically stable nanoemulsion. While at intermediate insoluble oil molar fractions, transient behavior was observed. However, the interfacial tension between oil and aqueous phase was not investigated in their study. In our case, the change pattern of particle size and interfacial tension was similar as Eu has a lower surface activity than the triglyceride. Hence, we induced that the difference in interfacial tension might be the main reason causing the change of particle size. As shown in Fig. 5B, 10 wt% ≤ X₂ ≤ 30 wt%, nanoemulsions were stable. At 30 wt% < X₂ < 70 wt%, the nanoemulsions were unstable and then at X₂ ≥ 80 wt%, the nanoemulsions were kinetically stable. The stability mechanism of the Eu emulsion containing a high triglyceride content was subsequently presented in Fig. 6. There was no coalescence among the emulsion droplets. Ostwald ripening was the main instability mechanism. The results illustrate that the change in interfacial tension needs to be taken into account for the cutoff of Ostwald ripening.

Fig. 6 Schematic stability mechanism of the eugenol emulsion in the presence of a high triglyceride content.

5. Conclusions

The presented work investigated the physical stability of the Eu nanoemulsions in the presence of triglycerides with a high incorporation ratio in the organic phase. When the triglyceride content changed from 30 wt% to 100 wt%, the particle size of the Eu nanoemulsions increased at first and then decreased. The Eu nanoemulsions had the smallest particle size and maximum instability index in the presence of 60 wt% triglyceride. Meanwhile, the particle size increased upon longer storage time. Especially with the incorporation of 60 wt% BO, phase separation was found after a five-month storage or after undergoing a centrifugal force. In contrast, when MCT was added, Eu nanoemulsions remained visually homogeneous with a slower increase in the particle size. The change in the interfacial tension as a function of triglyceride content exhibited the same trend as that of particle size. The results demonstrated that the interfacial tension and type of surfactant needed to be taken into account to fabricate stable essential oil nanoemulsions.

Conflict of interest

The authors have no declared conflict of interest.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (No. 31401528), National Science-technology Support Plan Projects (No. 2014BAK19B03) and China Postdoctoral Science Foundation (No. 2015M582237).

References

1. T. Koeduka, S. Suzuki, Y. Iijima, T. Ohnishi, H. Suzuki, B. Watanabe, D. Shibata, T. Umezawa, E. Pichersky and J. Hiratake, Biochem. Biophys. Res. Commun., 2013, 436, 73–78.
2. S. K. Roberts, M. McAinsh, H. Cantopher and S. Sandison, PLoS One, 2014, 9, e102712.
3. K. P. Devi, S. A. Nisha, R. Sakthivel and S. K. Pandian, J. Ethnopharmacol., 2010, 130, 107–115.
4. P. K. Lai and J. Roy, Curr. Med. Chem., 2004, 11, 1451–1460.
5. D. J. McClements, in Annual Review of Food Science and Technology, Annual Reviews, ed. M. P. Doyle and T. R. Klaenhammer, Palo Alto, 2010, vol. 1, pp. 241–269.
6. D. T. Piorkowski and D. J. McClements, Food Hydrocolloids, 2014, 42, 5–41.
7. F. Niu, W. Pan, Y. Su and Y. Yang, Food Chem., 2016, 212, 138–145.
8. L. Salvia-Trujillo, A. Rojas-Graü, R. Soliva-Fortuny and O. Martín-Belloso, Food Hydrocolloids, 2015, 43, 547–556.
9. Y. Chang, L. McLandsborough and D. J. McClements, J. Agric. Food Chem., 2012, 60, 12056–12063.
10. D. J. McClements, L. Henson, L. M. Popplewell, E. A. Decker and S. J. Choi, J. Food Sci., 2012, 77, C33–C38.
11. Y. Chang, L. McLandsborough and D. J. McClements, Food Chem., 2015, 172c, 298–304.
12. N. Terjung, M. Loffler, M. Gibis, J. Hinrichs and J. Weiss, Food Funct., 2012, 3, 290–301.
13. W. Li, H. Chen, Z. He, C. Han, S. Liu and Y. Li, LWT–Food Sci. Technol., 2015, 62, 39–47.
14. E. Nazarzadeh, T. Anthonypillai and S. Sajjadi, J. Colloid Interface Sci., 2013, 397, 154–162.
15. T. J. Wooster, M. Golding and P. Sanguansri, Langmuir, 2008, 24, 12758–12765.
16. R. Liang, S. Xu, C. F. Shoemaker, Y. Li, F. Zhong and Q. Huang, J. Agric. Food Chem., 2012, 60, 7548–7555.
17. Y. Li, S. Le Maux, H. Xiao and D. J. McClements, J. Agric. Food Chem., 2009, 57, 9243–9249.
18. K. Ziani, Y. H. Chang, L. McLandsborough and D. J. McClements, J. Agric. Food Chem., 2011, 59, 6247–6255.
19. F. Yuan, D. Xu, X. Qi, J. Zhao and Y. Gao, Food Bioprocess Technol., 2011, 6, 1024–1031.
20. S. Suriyarak and J. Weiss, Colloids Surf., A, 2014, 446, 71–79.
21. P. Taylor, Adv. Colloid Interface Sci., 1998, 75, 107–163.
22. A. S. Kabalnov, J. Dispersion Sci. Technol., 2001, 22, 1–12.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra16270c

‡ Co-first author with the same contribution to this work.

---